A server rack can include a plurality of nodes (e.g., rackservers). In order to prevent the plurality of nodes from overheating and failing, the nodes can be cooled by a shared fan cooling system. Shared fan cooling system attempt to cool the plurality of nodes as a single unit. That is, a server rack temperature is taken and the shared fan cooling system operates based on the server rack temperature. However, each node has an independent temperature. The independent temperature of a single node can be affected by the node's operating capacity. Nodes can operate at different capacities. The capacities can be based on CPU utilization. For example, a node can be idle (i.e., less than 2 percent processing capacity used), light (i.e., between 2 and 30 percent processing capacity used), medium (i.e., between 31 and 65 percent processing capacity used), or heavy (i.e., between 66 and 100 percent capacity used). Thus, a shared fan cooling system that takes each node into consideration is needed.